NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Obsreports > FocusProblemApril2005

Changes | Index | Notify | Search | Go

GBT Subreflector Problem April 3rd through 7th 2005

Overview

Starting after the power outage on Sunday April 4rd, the antenna control computer (Virgo) was not correctly synchronized to our site clock. This problem was cured on the afternoon of Thursday 7th. A result of this loss of synchronization was, on occasion, errors in the positions and tilts of the remaining axes when the subreflector was scanned in Y to perform a "focus" measurement. This problem was first noticed by Karen O'Neil, supporting GBT05A_007_02, for which the Z-tilt of the subreflector was so large that the telescope beam missed the source completely.

Due to the nature of the fault, we believe only actual focus scans would be affected. The subreflector would servo correctly when simply tracking a source. Further, the severity of the fault would change on an ~15 minute timescale.

I have spot checked the astronomical data taken between April 3rd and 7th, and indeed it seems that all project have at least one bad focus. I've also confirmed that this problem does appears to be directly related to the power outage, and was not occuring before April 3rd.

Symptoms

Again, non-focus data would not be affected by the fault. However, if an "incorrect" focus fit was accepted, subsequent data would have been acquired with a non-optimal focus.

How can I tell if I had a problem?

Look at your focus fits. If the peak height was similar to the preceding peak, the fitted FWHM close to expected, and the focus offset "reasonable", you were probably ok. If any of the above do not apply, the focus measurement was probably invalid. If you suspect a problem, look at the ANTPOSGR binary table for the Antenna FITS file of the scan in question. Look at the columns SR_XP through SR_ZT respectively (six values). I find it most convenient to plot them versus row number.

When the subreflector is behaving correctly, XP and ZP should show ~linear changes (the focus tracking curve) with an rms acatter of around 0.02mm. YP will show the focus scan. All of XT, YT and ZT should have fixed values, with rms scatters around 0.005 degrees. If the subreflector was suffering from the fault, the plots will not look like this. I've attached below comparion plots for a "good" scan (scan 28) and an extreme "bad" scan (scan 26) below. In this case, the ZT value (last plot) was clearly way off the expected value.

If anyone wants advice on whether specific focus curves are good or not, I'd be happy to assist.

What if I suspect this problem may have reappeared?

To check the antenna control computer's current time synchonization, run the command (from any host):

ntpq -c pe virgo

and you should see:

     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
+yed.gb.nrao.edu .PPS.            1 u    1   16  377    1.030    0.007   1.001
oPPS(1)          .PPS.            1 l    2   16  377    0.000   -0.007   0.031

The important columns are 'offset' and 'jitter'. Both values should be in the range -20 to +20, zero being optimal.

Plots

Scan 28 is typical of a "good" scan. Scan 26 is a rather extreme bad scan.

PDF versions

• 28 X position (SR_XP)
• 26 X position (SR_XP)
• 28 Y position (SR_YP)
• 26 Y position (SR_YP)
• 28 Z position (SR_ZP)
• 26 Z position (SR_ZP)
• 28 X tilt (SR_XT)
• 26 X tilt (SR_XT)
• 28 Y tilt (SR_YT)
• 26 Y tilt (SR_YT)
• 28 Z tilt (SR_ZT)
• 26 Z tilt (SR_ZT)

PNG versions